Methodology for Generating Ensemble Stream Discharge Forecasts to Feed an Ensemble Flood Forecasting System in a Coastal Flood Plain

The overarching objective is to predict inundation occurring from storms and flooding events using the New York Harbor Observing and Prediction System (NYHOPS). NYHOPS was developed at Stevens Institute of Technology's Davidson Laboratory to provide an “eye” on the Atlantic/ Hudson region through in-situ monitoring equipment and hydrodynamic modeling. One of the main inputs for the NYHOPS hydrodynamic component is freshwater discharge from inland hydrology. To account for such inputs, a state-of-the art stream discharge forecasting framework was implemented using gridded precipitation data from ensemble-based weather forecast models. In total, 66 precipitation datasets were used in a hydrologic model (HEC-HMS) and this includes the North American Mesoscale Forecast System (NAM), the Global Ensemble Forecast System (GEFS), the Short-Range Ensemble Forecast (SREF) and the Canadian Meteorological Centre (CMC).

The operational platform was applied to the Hudson River Basin (14,000 sq mi) which originates from the Adirondack Mountains of Upstate New York and drains into the Atlantic Ocean, between New York City and Jersey City. The framework is fully automated on Linux and produces 4 forecast cycles of ensemble river discharge per day that feed into NYHOPS in order to forecast floods for up to 72 hours out.

The results of this work demonstrate that ensemble stream discharge forecasts provide improved predictions and useful information about associated uncertainties, thus improving the assessment of risks as when compared to a single model forecast.